Influences of Animal Pollination and Seed Dispersal on Winter Flowering in a Temperate Mistletoe

Ecology, 84(10), 2003, pp. 2613±2627 ᭧ 2003 by the Ecological Society of America

INFLUENCES OF ANIMAL POLLINATION AND SEED DISPERSAL ON WINTER FLOWERING IN A TEMPERATE MISTLETOE

MARCELO A. AIZEN1 Laboratorio Ecotono, CRUB, Universidad Nacional del Comahue, Unidad Postal Universidad, 8400 Bariloche, RõÂo Negro, Argentina

Abstract. The evolution of ¯owering time can be in¯uenced directly by pollinators and indirectly by seed dispersers. In temperate latitudes, interactions with climate, especially temperature, may affect both plants and their animal mutualists. Winter-¯owering plants allow assessment of biotic in¯uences on the evolution of ¯owering time because temperature is unlikely to select directly for this uncommon reproductive behavior. In northwestern Patagonia, the hemiparasitic mistletoe Tristerix corymbosus ¯owers from early fall, through the cool winter, to late spring. This species is pollinated by the hummingbird Sephanoides galeritus, and its seeds are dispersed by the marsupial Dromiciops australis. In two populations during two years, I analyzed seasonal variation in pollination, fruit production, and fruit removal. Hummingbird visitation was lowest in winter and late spring, and ¯owers opening during those periods showed reduced pollination and fruit set, partly due to pollinator limitation, compared to ¯owers that opened during fall or early spring. Fruits that ripened during summer (January±March) had a higher chance of being removed than fruits ripening during either spring or fall, due to their overlap with the period of maximum disperser activity. Timing of ¯ower opening was strongly associated with fruit maturation time and with fruit removal rates. Thus, even though ¯owers that opened during winter exhibited only moderate fruit production, their fruits bene®ted from high fruit removal and seed dispersal in mid to late summer. These results suggest that the activity period of this plant's disperser, in combination with a lack of strong pollination constraints, probably governed the evolution of ¯owering phenology in this mistletoe. However, the proximate in¯uence of temperature on ¯ower and fruit development may prevent a ®ne-tuned match between ¯owering phenology and the period of maximum mutualist activity. Key words: ¯owering phenology; fruit removal; fruit set; hummingbird pollination; marsupial dispersal; mistletoe; pollen limitation; seed dispersal; Tristerix corymbosus; winter ¯owering.

INTRODUCTION speci®c pollination mutualisms for which reproductive fate depends largely on the temporal overlap between The evolution of ¯owering time, its duration, and ¯owering and the activity period of one or a few pol- synchrony may be affected by the interactions of plants linators (Rathcke and Lacey 1985, Bronstein 1995). In with their pollinators and seed dispersers (Stiles 1977, plant species sharing pollinators, individuals can also Wheelwright 1985, Bronstein 1995, Stone et al. 1998). increase their reproductive success by ¯owering when However, these two types of mutualisms in¯uence the plants of other species do not, thus avoiding compe- evolution of ¯owering phenology through different tition for pollinator service (Rathcke 1988, Feinsinger paths. Most obviously, patterns of pollinator activity et al. 1991, Stone et al. 1998). may directly promote a particular time and duration of Fruiting follows ¯owering and, at least in some spe- ¯owering. On the other hand, the in¯uence of seed cies, the duration of the fruit ripening period may be dispersers on ¯owering phenology is always indirect genetically ®xed or constrained by fruit size (Rathcke and mediated by their effect on fruiting time. and Lacey 1985, Primack 1987). Hence selection for Pollinator abundance typically ¯uctuates over the fruit maturation during periods of high disperser avail- potential ¯owering season. Thus, plants can enhance ability may indirectly in¯uence ¯owering phenology their reproductive success by opening ¯owers during and even relegate ¯ower opening to a period of sub- periods of high pollinator activity, which will ensure optimal climatic conditions for pollinator activity high pollen removal and export to other ¯owers as well as high pollen import (Mazer et al. 1989, Harder and (Wheelwright 1985). Wilson 1994, Aizen 2001). Such direct selection could The direct or indirect in¯uences of plant±animal mu- be particularly strong for plant species engaged in very tualisms on ¯owering phenology may be most apparent in tropical latitudes (Stiles 1977, Wheelwright 1985, Stone et al. 1998). While such in¯uences can also occur Manuscript received 22 August 2002; revised 14 January 2003; accepted 23 January 2003; ®nal version received 18 Feb- in temperate regions, they may be more dif®cult to ruary 2003. Corresponding Editor: K. D. Woods. disentangle from abiotic factors. In particular, milder 1 E-mail: [email protected] temperatures during spring and summer can be favor- 2613 2614 MARCELO A. AIZEN Ecology, Vol. 84, No. 10 able for both ¯ower development (e.g., Johnson 1992, America, I predicted higher hummingbird visitation to Lechowicz 1995, Diekmann 1996) and the activity of T. corymbosus ¯owers that open during winter when animal mutualists (Rathcke and Lacey 1985). Thus, no alternative nectar resources are available. Elevated winter-¯owering plants are particularly suitable bio- visitation should induce higher pollination and seed logical models to test hypotheses about the role of bi- output for ¯owers opening during winter than either in otic interactions in shaping ¯owering phenology in the fall or spring. (H2) Winter ¯owering favors seed dis- temperate zone, because climate (and temperature in persal (``disperser-availability'' hypothesis). Because particular) is unlikely to be the primary selective factor the abundance of seed dispersers varies seasonally, I for such an unusual reproductive behavior. Despite predicted higher dispersal of fruits ripening from winter their taxonomic rarity, plant species that ¯ower during ¯owers than from either fall or spring ¯owers, if winter winter may play a critical ecological role in maintain- ¯owers mature fruit when dispersers are most abun- ing resident populations of key pollinators and frugi- dant. Implicit in this hypothesis is a strong correlation vores (Hopper 1981, Smith-RamõÂrez 1993, Armesto et between date of fruit maturation and date of ¯ower al. 1996). In turn, availability of pollinators and ef®- opening. cient seed dispersers could strongly in¯uence the timing of reproductive events in such plants (Rathcke and MATERIALS AND METHODS Lacey 1985, Primack 1987). The plant species and study area In this paper, I assess the possible in¯uences of plant±animal mutualisms (i.e., pollination and seed dis- Tristerix corymbosus (quintral) is a shrubby mistletoe persal) on the evolution of ¯owering phenology in a distributed along the Paci®c rim of South America from winter-¯owering plant. In particular, I investigated the 32Њ to 42Њ S (Kuijt 1988). Reproductive plants can pro- dependence of pollination, fruit set, and fruit removal duce several hundred ¯ower buds that differentiate and on the month of ¯ower opening in Tristerix corymbosus develop during summer. The tubular ¯owers are large (Loranthaceae), a hummingbird-pollinated and mar- (ϳ5 cm long), red and radially symmetrical, and clus- supial-dispersed mistletoe native to the temperate for- tered in in¯orescences with 4±14 ¯owers. The ®laments est of southern South America. This plant ¯owers from and style, which are yellow during anthesis, turn red March (austral fall) to November (austral spring), dur- when a ¯ower senesces (Tadey and Aizen 2001). Con- ing the cold and snowy winters that characterize the trolled pollination showed that this mistletoe is self- southern Andean part of its geographic range. compatible, but animal visitation is needed to achieve The temperate forest of southern South America full reproductive output (M. A. Aizen, unpublished hosts a rich hummingbird-pollinated ¯ora comparable data). Flowers mature into 1 ϫ 0.5-cm green pseudob- to that of many Neotropical forests (Aizen and Ezcurra erries, each containing one ``naked'' seed 0.7 cm long 1998, Aizen et al. 2002). Unlike hummingbird ¯oras surrounded by a sweet viscous pulp (Amico and Aizen from lower latitudes, this temperate ornithophilous ¯o- 2000). After a few weeks and if not removed, fruits ra is pollinated principally by a single species, Se- senesce on the plant. Senescent fruits are characterized phanoides galeritus (Cocucci 1991, Smith-Ramirez by a wrinkled, black pericarp and a bitter pulp. 1993, Armesto et al. 1996). Therefore, T. corymbosus The study was conducted in areas adjacent to Nahuel may bene®t from opening ¯owers during winter, when Huapi National Park, in the eastern foothills of the no other plants bloom, by monopolizing the foraging Patagonian Andes in southern Argentina. At 785 m, activity of its hummingbird pollinator. the altitude of Lake Nahuel Huapi, mean July and Jan- In this temperate forest, seeds of T. corymbosus are uary temperatures are 2.2ЊC and 14.1ЊC, respectively dispersed exclusively and ef®ciently by the endemic (1914±1996 temperature record at Bariloche airport). arboreal marsupial Dromiciops australis (Amico and A strong W±E precipitation gradient ranges from Aizen 2000). Therefore, even though winter ¯owering Ͼ3500 to Ͻ700 mm in less than 50 km from the con- could expose T. corymbosus to poor pollination due to tinental divide. On average, only 12% of the annual severe weather, it could also promote dispersal success precipitation falls during the summer months (Decem- if ¯owers that open during winter mature fruit during ber±February). Snowfalls are common during winter summer when disperser abundance peaks (Rau et al. (Barros et al. 1983). 1995, Amico 2000). This proposition assumes that ei- I studied ¯owering phenology, animal visitation to ther selection on fruit development time is weak or that ¯owers, pollination, fruit set, and fruit removal during this trait is highly canalized and unlikely to respond to two complete ¯owering±fruiting periods (1997±1998 selection. and 1998±1999) in two populations situated 10 km In this study, I took advantage of the extended phe- apart and ϳ25 km west of the city of San Carlos de nology of T. corymbosus to address the following hy- Bariloche (41Њ8Ј S, 71Њ19Ј W). One population is in potheses: (H1) Winter ¯owering increases pollination the Llao-Llao forest reserve (hereafter Llao-Llao), a and seed production (``pollinator-availability'' hypoth- mesic old-growth forest dominated by Nothofagus esis). Because no other hummingbird-pollinated plant dombeyi (mean annual precipitation ϳ1900 mm), and ¯owers during winter in the temperate forest of South the other is in the PenõÂnsula de San Pedro (hereafter October 2003 POLLINATION AND SEED DISPERSAL 2615

PenõÂnsula) on the southern margin of Lake Nahuel Hua- dition, I hand-pollinated half of all tagged ¯owers on pi (mean annual precipitation ϳ1400 mm). At Llao- all sampled plants rather than on a subset of plants. Llao, T. corymbosus mostly parasitizes branches of the Results of comparing plants with and without hand- shrub Aristotelia chilensis, whereas at PenõÂnsula it oc- pollinated ¯owers from the previous year showed that curs on the tree Maytenus boaria. open-pollinated ¯owers have the same probability of setting fruit, independent of the presence of hand-pol- Field sampling linated ¯owers (one-way ANOVA; F1,16 ϭ 0.16, P ϭ

During the 1997 ¯owering season (March±Decem- 0.69 and F1,13 ϭ 0.05, P ϭ 0.83 for fruit set at Llao- ber) and subsequent 1997/1998 fruiting season (Oc- Llao and PenõÂnsula, respectively). Thus, an increase in tober±April), I sampled ¯owers and fruits in 20 indi- fruit set among hand-pollinated T. corymbosus ¯owers viduals of T. corymbosus in each of the two popula- can be attributed to pollen limitation rather than trans- tions. At weekly intervals, I counted the number of location of resources from lightly to heavily pollinated open ¯owers on three tagged branches per plant. Also, ¯owers (see Fox 1992). Over both years, I acquired I marked 3±5 swollen ¯ower buds per plant per week data on ¯owering, pollination, and fruiting for ϳ10 000 using numbered jewelry tags. Thereafter, I checked the ¯owers. status of each tagged bud at weekly intervals and classi®ed it according to one of the following categories: Data analysis bud, open ¯ower, senescent ¯ower, swollen ovary, de- To address how pollination and seed dispersal might veloping fruit, mature fruit, and senescent fruit. On have in¯uenced the evolution of ¯owering phenology each sampling date, I also recorded the number of hum- in T. corymbosus, I used a ¯ower-cohort approach rath- mingbirds or other ¯ower visitors observed foraging er than the more common phenotypic-selection ap- on T. corymbosus while I was in the ®eld. For even- proach (e.g., WideÂn 1991, GoÂmez 1993, O'Neil 1999), numbered plants (i.e., 10 individuals per population), because the ¯owering phenology of this mistletoe I hand-pollinated half of all tagged ¯owers at anthesis varies little among plants (see Results). Although this using pollen collected from 3±5 other individuals to approach does not permit estimation of the strength of test for pollen limitation of fruit set. natural selection, it allows assessment of the repro- To assess pollination, I collected styles from ap- ductive bene®ts associated with ¯owers opening at dif- proximately half of all tagged ¯owers (either open- or ferent times and identi®cation of potential factors in- hand-pollinated) ϳ2±3 wk after ¯owers were ®rst re- ¯uencing the evolution of ¯owering phenology (e.g., corded as senescent. Styles were kept individually in Pico and Retana 2000, Wolfe and Burns 2001). Here, microcentrifuge tubes containing formalin:acetic acid: I classi®ed and averaged data by month of either ¯ower ethyl alcohol (FAA, 5:5:90). In the laboratory, styles opening (i.e., March, April, May, etc.) or fruit matu- were cleared overnight in 10 mol/L NaOH, stained with ration (i.e., October, November, December, etc.).

0.1% aniline blue in 0.1 mol/L K3PO4, squashed, and Flowering phenology.ÐI described ¯owering phe- examined with an epi¯uorescence microscope (Martin nology as the proportion of all ¯owers counted on 1959). I later counted the number of pollen tubes pen- tagged branches that were open during a given ¯ow- etrating the stigma (i.e., just below the stigma surface). ering month. Monthly values were then averaged across Fruits of T. corymbosus remain green when ripe. individuals. However, because the number of open Therefore, I considered a fruit to be mature when it ¯owers re¯ects both ¯ower opening rate and ¯ower was turgid, but soft to the touch. Approximately 10% longevity (Primack 1985), ¯ower counts may not mea- of all fruits that developed from tagged ¯owers were sure accurately the number of ¯owers opening during bagged with nylon mesh just before maturation to es- a period if ¯ower longevity varies seasonally. There- timate maximum retention time. Based on observations fore, for each individual I estimated monthly means of from bagged fruits, mature fruits rarely fall naturally ¯ower longevity (i.e., time a ¯ower remained open) in (ϳ2%). Thus I considered a fruit to have been removed weeks (individual ¯owers last 0±4 wk). Then I esti- by a disperser when an exposed fruit that had been mated the proportion of all ¯owers coming into anthesis classi®ed in a given census as mature was absent in in a given month by weighting the number of open the following census. The marsupial disperser of T. ¯owers by the inverse of mean ¯ower longevity for corymbosus does not remove or consume senescent that month. Resulting values were restandardized (thus fruits (personal observation), so that when a senescent the sum of this variable over ¯owering months equaled fruit disappeared between censuses it was considered one) and then averaged across individuals. as fallen. The pollinator-availability hypothesis.ÐI computed I sampled ¯owers and fruits from the same individ- a relative visitation index as the number of humming- uals during the following ¯owering and fruiting season birds (or other ¯ower visitors) observed foraging dur- (i.e., 1998 and 1998/1999), except that one plant in ing a given sampling date divided by the product of Llao-Llao and six plants in PenõÂnsula died and were number of hours I spent in the ®eld and the number of replaced. Sample sizes during this second year totaled open ¯owers on tagged branches for that date (i.e., 23 plants (Llao-Llao) and 21 plants (Peninsula). In ad- number of hummingbirds per hour per ¯ower). I av- 2616 MARCELO A. AIZEN Ecology, Vol. 84, No. 10 eraged visitation index values by month for each site I analyzed seasonal trends in fruit removal using and year. Although this index cannot be used to com- ANOVA with month of fruit ripening (or alternatively pare visitation to ¯owers between sites or years because month of ¯ower opening) as the main factor and in- I counted ¯owers in a subset of branches and plants dividual plant as a blocking factor. I also used ANOVA occurring in each population, it should provide an un- to assess the effects of fruit bagging, month of fruit biased estimate of the seasonal change in pollinator ripening (or alternatively month of ¯ower opening), activity at T. corymbosus ¯owers. and the bagging treatment ϫ month interaction on mean For each plant and month of ¯owering, mean number fruit retention time. According to the disperser avail- of pollen tubes and fruit set were calculated for open- ability hypothesis (H2), fruit removal should increase and hand-pollinated ¯owers separately. I assessed sea- and fruit retention time should decrease among fruits sonal trends in pollination and fruit set and temporal set by ¯owers opening during winter months. changes in pollen limitation using ANOVA with pol- Per-¯ower contribution to ®tness.ÐI estimated the lination treatment (i.e., open vs. supplemental hand- cumulative success of a ¯ower opening in a given pollination) and month of ¯owering as crossed main month by integrating pollination and seed dispersal into factors and individual plant as a blocking factor. Ac- a single measure. I used fruit set as a measure of pol- cording to the pollinator-availability hypothesis (H1), lination success and fruit removal as a surrogate of seed differences in pollination and fruit set between open- dispersal. Therefore, the bene®ts of interacting with and hand-pollinated ¯owers should be smaller during mutualists for a ¯ower opening in month i (Fi) can be winter than either fall or spring months. summarized as Flower-to-fruit time interval and fruit phenology.Ð F ϭ fr rm (2) From date of individual ¯ower opening and fruit mat- iii uration, I estimated the time from ¯ower anthesis to where fri and rmi are the fruit set and removal rate, fruit. I also calculated the relative proportion of fruits respectively, exhibited by a ¯ower opening during that matured during a given month from different ¯ow- month i. This measure represents the probability that er cohorts. For a given site and year, I used a likelihood a ¯ower opening in a given month sets a fruit that is ratio test (G test) to compare the proportion of fruits removed and weighs the relative in¯uences of polli- that ripened each month originating from ¯owers that nation and seed dispersal on a ¯ower's contribution to opened in the fall (March±May), winter (June±August), plant ®tness. For each population and year, I analyzed and spring (September±December). seasonal variation in Fi using ANOVA with month of For each plant, I estimated the proportion of all fruits ¯owering as the main factor and individual plant as the that ripened during month m (i.e., frm)as blocking factor. fr ϭ fl fr p (1) miiim͸ RESULTS i Flowering phenology where ¯i is the proportion of all ¯owers opening during month i, fri is the proportion of those ¯owers setting In the study populations, ¯owering extended from fruit, and pim the proportion of those fruits that ripened March to December in 1997 and from March to No- during month m. Resulting values were then averaged vember in 1998 (Fig. 1). Individual plants ¯owered across individuals from a given population and year. throughout most of the ¯owering period for their pop- The disperser-availability hypothesis.ÐFor each ulation. Nineteen of 20 surviving plants at Llao-Llao plant, I estimated probabilities of fruit removal by and 17 of 17 at PenõÂnsula ¯owered uninterruptedly for month of fruit ripening and by month of ¯ower opening eight or more months in the 1997 season. Comparable based on the proportion of fruits that ripened in a given ®gures for 1998 were 17 of 23 and 15 of 20 individuals. month (or alternatively produced by ¯owers that The onset of ¯owering was highly synchronous. During opened in a given month) that I considered as removed 1997 all sampled individuals started ¯owering in by an animal vector. Fruit removal can be used as a March. During 1998, at Peninsula, 19 individuals start- surrogate for seed dispersal for T. corymbosus because ed ¯owering in March and one in April; at Llao-Llao the marsupial Dromiciops australis effectively dis- 14 individuals started ¯owering during March, ®ve dur- perses mistletoe seeds and is the main frugivore that ing April, and four during May. The number of open removes and ingests fruits in the study populations of ¯owers peaked during May±June in 1997 at both sites T. corymbosus (see Discussion). and in 1998 at PenõÂnsula. Flowering at Llao-Llao in I also computed monthly means of total fruit reten- 1998 peaked during August (Fig. 1). I counted 45±48% tion time for exposed fruits (i.e., available for removal) of all ¯owers on tagged branches during winter (June± and bagged fruits separately. I estimated retention times August). of individual fruits as the number of weeks between Flower longevity averaged slightly more than 1 wk the time a fruit was ®rst considered mature and the (1.14 Ϯ 0.016 wk, Llao-Llao 1997; 1.04 Ϯ 0.016 wk, time it was recorded as absent (either ``removed'' or PenõÂnsula 1997; 1.15 Ϯ 0.015 wk, Llao-Llao 1998; and ``fallen''). 1.05 Ϯ 0.015 wk, PenõÂnsula 1998; means Ϯ 1 SE), but October 2003 POLLINATION AND SEED DISPERSAL 2617

FIG. 1. Flowering phenology of Tristerix corymbosus in the Llao-Llao and PenõÂnsula sites in areas adjacent to Nahuel Huapi National Park, in the eastern foothills of the Patagonian Andes in southern Argentina during 1997 and 1998. Bars depict the percentage of all ¯owers observed open in a given month (means Ϯ 1 SE). Dotted lines depict estimates of ¯ower longevity. Solid lines indicate the percentage of all ¯owers coming into anthesis in a given month. increased during winter (Fig. 1). The mean ¯ower open- ited mistletoe ¯owers during early fall at PenõÂnsula and ing in July lasted 2±3 times longer than ¯owers opening during late spring at Llao-Llao in both years. Hum- in March or November. Seasonal differences in ¯ower mingbirds visited ¯owers throughout the ¯owering sea- longevity were more marked in 1997 (Fig. 1). son with maximum visitation in early fall and declining Because of the seasonal change in ¯ower longevity, to a minimum during late spring. A secondary peak in the phenological patterns in counts of open ¯owers hummingbird visitation occurred in September. differed from the patterns in date of ¯ower opening: In both populations and years, the number of pollen the proportion of ¯owers entering anthesis in a given tubes per style showed signi®cant seasonal variation month peaked earlier (April) in both populations in (Appendix A). In open-pollinated ¯owers, pollination 1997 and at PenõÂnsula in 1998 (Fig. 1). Only 29% and peaked at the beginning of the ¯owering season 26% of all ¯owers produced during 1997 and 38% and (March) and again during September 1997 and July± 33% during 1998 at Llao-Llao and PenõÂnsula, respec- August 1998 (Fig. 3). With one exception (PenõÂnsula tively, entered anthesis during winter months. Thus, 1997), May- and June-opening ¯owers had lower num- the full-bloom winter appearance of this mistletoe is accentuated by ¯ower persistence. bers of pollen tubes in open-pollinated ¯owers. In both sites and years, pollination levels in open-pollinated Flower visitation, pollination, and fruit set ¯owers declined strongly by the end of the ¯owering The hummingbird Sephanoides galeritus was the season (October±December). Styles of open-pollinated most frequent ¯ower visitor, accounting for Ͼ85% of ¯owers had, on average, signi®cantly fewer pollen all individuals observed visiting ¯owers of T. corym- tubes than hand-pollinated ¯owers (Appendices A and bosus (Fig. 2). Queens of Bombus dahlbomii also vis- B). 2618 MARCELO A. AIZEN Ecology, Vol. 84, No. 10

FIG. 2. Seasonal changes in relative visitation rates (visitation index) to the ¯owers of Tristerix corymbosus in Llao-Llao and PenõÂnsula over 1997 and 1998 by hummingbirds and bumble bees. Values are means Ϯ 1 SE.

With one exception (PenõÂnsula 1998), there were no hand-pollinated ¯owers set signi®cantly more fruit than signi®cant interactions between pollination treatment open-pollinated ¯owers (Appendices A and B). and month of ¯ower opening (Appendix A). However, Although interacting effects of month and treatment examination of Fig. 3 shows some consistent differ- on fruit set were signi®cant and marginally signi®cant ences in the seasonal variation exhibited by open- and only in Llao-Llao 1997 and PenõÂnsula 1998 (Appendix hand-pollinated ¯owers, particularly pronounced in A), pollination limitation was more common during ¯owers opening during May 1998 (both populations) some months than others (Fig. 4). With some excep- and during July 1998 (Llao-Llao). The decline in pol- tions, fruit production was more pollen-limited in ¯ow- lination in open-pollinated ¯owers entering anthesis at ers that opened in late fall to mid-winter (May±July) the end of the ¯owering season was not observed in and late in the ¯owering season (October±December). hand-pollinated ¯owers and thus, differences in pollination between treatments was greater for ¯owers that Flower-to-fruit time interval and fruit phenology opened in late spring in both years and populations Fruit development time showed strong seasonal pat- (Fig. 3). terns. The period from ¯ower anthesis to fruit matu- Fruit set showed highly signi®cant seasonal variation ration decreased over ¯owering at a rate of ϳ2 wk/mo (Appendix A). In parallel with pollination (Fig. 3), fruit from Ͼ30 wk for ¯owers opening early in the season set from open-pollinated ¯owers was highest for ¯ow- (March±April) to Ͻ20 wk (November±December) at ers opening at the beginning of the ¯owering season the end of the ¯owering season (Fig. 5). Despite de- (March±April) and in September (Fig. 4), whereas creasing maturation times, early ¯owers matured early ¯owers opening during early to mid-winter (June±July) fruits. Season of ¯ower opening was strongly associ- showed lower fruit set. Lower fruit set was also ob- ated with month of fruit ripening (G test, P Ͻ 0.0001 served among open-pollinated ¯owers opening after for both populations and years). Most fruits developing September (Fig. 4). In both populations and years, from fall ¯owers matured in spring (October±Decem- October 2003 POLLINATION AND SEED DISPERSAL 2619

FIG. 3. Seasonal changes in pollination of open- and hand-pollinated ¯owers at Llao-Llao and PenõÂnsula during the 1997 and 1998 ¯owering periods (means Ϯ 1 SE). Asterisks indicate individual months showing signi®cant differences (paired t test, P Ͻ 0.05) in number of pollen tubes between open- and hand-pollinated ¯owers after application of Bonferroni's correction for multiple tests (Rice 1989). See Appendix A for statistical analyses. ber), fruits from winter ¯owers matured in summer B and C). However, the proportion of fruits removed (January±February), and those from spring ¯owers ma- and the duration of fruit retention on the plants varied tured in late summer±early fall (March±May; Fig. 6). strongly according to month of fruit ripening (Fig. 7, Fruits developed more slowly during 1997, particularly Appendix C). In all cases, fruits maturing during sum- for ¯owers opening during the ®rst half of the ¯owering mer (January±March) had the highest probabilities of season (March±July; Fig. 5). As a result, a higher pro- removal and shortest retention times (Fig. 7). Fruits portion of fruits set from fall ¯owers delayed ripening maturing early in the season (October±November) until the following summer during 1997 than during showed low removal rates (Ͻ40%), and their mean 1998. Whereas 1997 fall ¯owers contributed 40.3% retention time did not differ from that of isolated, (Llao-Llao) and 24.7% (Peninsula) of all fruits ripening bagged fruits. Also, fruit removal declined by the end during the following January, these contributions de- of the fruiting season (April±May), particularly in creased to 11.3% and 10.9% for fruits set by the 1998 1997/1998 (Fig. 7). fall ¯owers (Fig. 6). Fruit ripening (Eq. 1) peaked in Because of the relationship between date of anthesis December, except at Llao-Llao in 1998/1999, which and date of fruit ripening, fruits developing from win- peaked in February (Fig. 6). ter-opening ¯owers (June±September) had higher removal rates (Ͼ70% and up to ϳ90%) and lower re- Fruit removal and retention time tention times (Ͻ3 wk) than fruits maturing from ¯owers Frugivores removed a large proportion of fruits (0.79 that open at the beginning and end of the ¯owering Ϯ 0.022, Llao-Llao 1997; 0.65 Ϯ 0.026, PenõÂnsula season (Appendix D, Fig. 8). Fruits set by ¯owers open- 1997; 0.61 Ϯ 0.018, Llao-Llao 1998; and 0.58 Ϯ 0.023, ing at the tails of the ¯owering phenology exhibited PenõÂnsula 1998). Exposed fruits were retained on the removal rates as low as ϳ20% and their retention times plants for a shorter time than bagged fruits (Appendices did not differ signi®cantly from bagged fruits. 2620 MARCELO A. AIZEN Ecology, Vol. 84, No. 10

FIG. 4. Fruit set of open- and hand-pollinated ¯owers in relation to month of ¯ower opening at Llao-Llao and PenõÂnsula during the 1997 and 1998 ¯owering phenologies (means Ϯ 1 SE). Asterisks indicate individual months showing signi®cant differences (paired t test, P Ͻ 0.05) in fruit set between open- and hand-pollinated ¯owers after application of Bonferroni's correction for multiple tests (Rice 1989). See Appendix A for statistical analyses.

Per-¯ower contribution to ®tness snowfalls (personal observation) and are effective pol- Monthly estimates of per-¯ower contribution to ®t- linators (Smith-RamõÂrez 1993, Tadey and Aizen 2001). ness (Eq. 2) varied over ¯owering season consistently In comparison, queens of Bombus dahlbomii visited across years and populations (Fig. 9). The probability ¯owers of T. corymbosus less frequently, mostly under of a ¯ower setting a removed fruit increased over ¯ow- relatively benign climatic conditions at either end of ering period, peaking during late winter (August±Sep- the ¯owering period (Fig. 2), often missing contact with anthers and stigmas (personal observation). tember) and decreasing abruptly afterwards. This pat- The relatively poor pollination of winter ¯owers tern resulted because ¯owers opening during late win- strongly suggests that pollination advantages cannot se- ter exhibited both high fruit set (Fig. 4) and a high lectively promote the odd ¯owering phenology exhibited chance of fruit removal when ripe (Fig. 8). Fall ¯owers by this mistletoe. According to the pollinator-availability (March±May) had a lower individual contribution to hypothesis (H1), winter ¯owers of T. corymbosus would overall plant ®tness despite high fruit set because of monopolize the hummingbird's attention in the absence poor removal. Flowers opening during mid- to late of alternative ¯oral resources, leading to higher polli- spring (October±December) contributed little to ®tness nation and seed set for winter ¯owers. Overall, I did not because of both poor fruit set and removal. ®nd support for this hypothesis. Contrary to expectation, DISCUSSION hummingbird visitation was highest during early fall, at the beginning of the ¯owering period of T. corymbosus, Winter ¯owering and pollination and declined strongly in winter with a secondary peak Sephanoides galeritus was the most important, if not during early spring. Accordingly, ¯owers opening dur- exclusive, pollinator of T. corymbosus, interacting with ing early winter (June) and mid- to late spring (October± this mistletoe throughout its ¯owering phenology. December) had lower pollination and fruit set than those Hummingbirds visited ¯owers even during heavy that opened during early fall (March) or early spring October 2003 POLLINATION AND SEED DISPERSAL 2621

rates of hummingbird visitation. First, despite the re- quirement of hummingbird visits for maximum seed set, this plant exhibits a limited capacity for autonomous self-pollination (about one-third of full seed set; M. A. Aizen, unpublished data). Autonomous within- ¯ower self-pollination is a common feature of the breeding system of many winter and early-spring ¯owering plants (Schemske et al. 1978, Motten 1982, Her- rera et al. 2001). Second, particularly for ¯owers opening during winter, prolonged ¯ower longevity could partially offset dwindling hummingbird visitation rates (cf. Primack 1985, Totland 1994). Winter temperatures may also affect pollination and fruit set, independent of pollinator activity or other buffering factors. In this study, evidence of environmental stress on plant sexual reproduction can be in- ferred from the decline in the number of pollen tubes and fruit set in hand-pollinated ¯owers during one or

FIG. 5. Period from ¯ower anthesis to fruit maturation vs. more winter months (Figs. 3 and 4). Pollen grain ger- month of ¯ower opening (means Ϯ 1 SE). Observed curves mination, tube growth, and fruit development are all were ®tted by the following linear regression equations: y ϭ susceptible to low temperatures (Primack 1987, Ste- 2 44.6Ϫ2.44x (r ϭ 0.98, n ϭ 10) for Llao-Llao 1997; y ϭ phenson et al. 1992, Totland 1994). In addition, al- 41.2Ϫ2.08x (r2 ϭ 0.97, n ϭ 10) for PenõÂnsula 1997; y ϭ 39.8±1.94x (r2 ϭ 0.98, n ϭ 9) for Llao-Llao 1998; and y ϭ though ¯owers of T. corymbosus can withstand subzero 39.3Ϫ1.93x (r2 ϭ 0.97, n ϭ 9) for PenõÂnsula 1998. March temperatures, they can be frost-damaged when exposed was month 3 of the year. In all instances, P Ͻ 0.0001. to rapid temperature ¯uctuations (personal observation).

(September). Reduced fruit set from ¯owers opening Winter ¯owering and seed dispersal during winter and mid- to late spring could be ascribed Because of its aerial parasitic lifestyle, seed dispersal in part to limited pollen deposition. Consistent seasonal is critical for mistletoe survival. Dispersal of mistletoe patterns of variation in pollinator limitation over a spe- seeds involves not only the transport of seeds away ciesЈ ¯owering phenology has been reported for at least from the maternal plant, but also their placement on two other plant taxa (Ramsey 1995, O'Neil 1999). How- the living branches of an appropriate host. Given these ever, the lack of pollen limitation among winter ¯owers stringent requirements, it is not surprising that mistle- in the PenõÂnsula population during 1997 and late-spring toe-disperser interactions are among the most special- ¯owers of the Llao-Llao population during 1998 (Figs. ized and ef®cient plant±frugivore interactions known 3 and 4) indicates that the seasonal outcome of this (Reid 1991). mistletoe±pollinator interaction varies among years and Dromiciops australis is highly ef®cient in dispersing populations. T. corymbosus seeds. Almost all of the seeds it ingests The observed pattern in pollinator visitation to the are defecated undamaged, and Ͼ80% of them are ¯owers of T. corymbosus probably re¯ects seasonal placed on branches or trunks of host plants. Further- changes of hummingbird abundance at a larger spatial more, the passage of the seeds through the marsupial's scale, rather than changes in visitation rates associated gut is critical for germination and host infection (Ami- with variation in ¯ower nectar production (Rathcke co and Aizen 2000). Therefore, behavior of this mam- 1992). Hummingbird visitation to an unlimited source mal could exert strong selection on reproductive traits of arti®cial nectar over a 4-yr period (Appendix E) of T. corymbosus, including phenology. In particular, exhibited a seasonal trend similar to that found for winter ¯owering should be favored if it enhances seed ¯owers of T. corymbosus. Low winter abundance of dispersal by increasing the overlap between fruiting hummingbirds may result from partial emigration west- and the activity period of the marsupial; i.e., the dis- wards and northwards and a shift towards a more in- perser-availability hypothesis (H2). I found strong sup- sectivorous diet while breeding during October±No- port for this hypothesis. vember (C. Smith-RamõÂrez, personal communication). Fruit removal varied strongly over the 6±7 mo fruit- Although pollination peaked in March±April and ing period of T. corymbosus, with fruits ripening during again in August±September, in parallel with humming- mid- to late summer experiencing the highest removal bird activity, pollination and fruit set did not show the (Fig. 7). During summer, D. australis becomes the same decreasing trend throughout the ¯owering period. dominant arboreal mammal in the forests studied. Two aspects of the reproductive biology of T. corym- Dromiciops australis is also one of the few mammals bosus could buffer sexual reproduction against low in southern South America with con®rmed hibernation 2622 MARCELO A. AIZEN Ecology, Vol. 84, No. 10

FIG. 6. Percentage of fruits that matured during different months that derived from fall (March±May), winter (June± August), and spring (September±December) ¯owers at Llao-Llao and PenõÂnsula (bars). Fruit phenology (solid line, means Ϯ 1 SE) is indicated as the percentage of all fruits produced that ripened during a given month (Eq. 1). The 1997/1998 fruiting period corresponds to the 1997 ¯owering period, whereas the 1998/1999 fruiting period corresponds to the 1998 ¯owering period.

(Jaksic 1997) and is mostly inactive from late fall to and number of established seedlings per host (M. Rod- early spring, at least in the Andean part of its range rõÂguez-Cabal, unpublished data). These observations (Kelt 1994, Rau et al. 1995). Amico (2000) captured indicate that fruit removal rates satisfactorily depict the individuals of D. australis beginning in December at outcome of this mistletoe±marsupial dispersal inter- Llao-Llao. Abundance and proportion of juveniles in- action from the plant's perspective. creased signi®cantly through summer until March and Even with a strong association between dispersal decreased markedly afterwards. Thus, the summer peak rates and fruit phenology, the seasonal activity pattern in fruit removal is correlated with the seasonal activity of an ef®cient disperser can selectively in¯uence the pattern of the marsupial. Also, seasonal patterns of timing of ¯owering only if ¯ower-to-fruit develop- marsupial activity and abundance closely re¯ect tem- mental time is genetically ®xed or under weak selection poral variation in rates of seed deposition on branches (cf. Endler 1986). Flower-to-fruit time intervals in T. of host plants (Amico 2000). An ongoing study as- corymbosus responded to temperature (see Ratchcke sessing spatial variation in mistletoe dispersal across and Lacey 1985, Primack 1987); ¯owers opening dur- six T. corymbosus populations showed that local abun- ing fall produce fruit that remains unripe through win- dance of D. australis is a good predictor (r2 Ͼ 0.75) ter, exhibiting a development period twice as long as of the rates of mistletoe fruit removal, and fruit removal that observed in spring ¯owers. In addition, at both is a good predictor of both seed deposition on branches study populations fruit development was, on average, October 2003 POLLINATION AND SEED DISPERSAL 2623

FIG. 7. Proportion of fruits removed (circles) and retention time (squares) of exposed fruits in relation to month of fruit maturation in the Llao-Llao and PenõÂnsula populations during the 1997/1998 (October±May) and 1998/1999 (October±April) fruiting periods (means Ϯ 1 SE). Stars indicate individual months showing signi®cant differences (paired t test, P Ͻ 0.05) in retention time between exposed and bagged fruits (not shown) after application of Bonferroni's correction for multiple tests (Rice 1989). See Appendix C for statistical analyses.

1±2 wk shorter for ¯owers opening during the unusu- may also constitute indirect agents of selection on ¯ow- ally warm winter of 1998 than during the winter of ering phenology (English-Loeb and Karban 1992, Bro- 1997 (Fig. 5, Appendix E). On the other hand, I found dy 1997, PicoÂ and Retana 2000). However, those study low among-plant variability in mean fruit maturation species have short periods from ¯ower to seed and, in times (CVs ϳ 5%), and a strong coupling between ¯ow- many instances, eggs of seed-eating larvae are laid di- er and fruit phenology arising because differences in rectly on the ¯owers. the date of ¯ower anthesis were not offset by reductions in ¯ower-to-fruit maturation times of the same mag- Adaptation, constraints, and evolutionary scenarios nitude (Fig. 5). These ®ndings might indicate some This study provides clear evidence that winter ¯ow- developmental canalization and low genetic variation ering enhances plant ®tness in T. corymbosus princi- underlying this trait in T. corymbosus (cf. Endler 1986). pally by increasing the number of removed and even- The unimodal association between removal rate and tually dispersed seeds during summer. However, pol- date of fruit ripening (Fig. 7) translated into an equiv- lination and fruit production also in¯uence variation in alent humped relationship between fruit removal and the contribution of individual ¯owers to plant ®tness date of ¯ower anthesis (Fig. 8). Therefore, high dis- over the ¯owering season. In particular, high fruit set persal during summer clearly favors winter ¯owering determines a peak associated with ¯owers that open in T. corymbosus. I do not know of comparable studies during late rather than early winter. This seasonal pat- assessing the consequences of ¯owering phenology for tern was similar among years and populations (Fig. 9) animal seed dispersal (or fruit removal) within a plant suggesting that plant±animal mutualisms exert consis- species. Other studies have found that seed predators tent selection on ¯owering phenology over space and 2624 MARCELO A. AIZEN Ecology, Vol. 84, No. 10

FIG. 8. Proportion of fruits removed (circles) and retention time (squares) of exposed fruits in relation to month of ¯ower opening at Llao-Llao and PenõÂnsula (means Ϯ 1 SE). Stars indicate individual months showing signi®cant differences (paired t test, P Ͻ 0.05) in retention time between exposed and bagged fruits (not shown) after application of Bonferroni's correction for multiple tests (Rice 1989). See Appendix D for statistical analyses.

time in T. corymbosus, a prerequisite for these inter- late-summer fruits are set mainly by late-winter and actions to result in evolutionary change (Herrera 1988, spring ¯owers, seedling establishment is likely to ac- Horvitz and Schemske 1990, Willson and Whelan centuate the contribution of winter ¯owers to plant ®t- 1993, Waser et al. 1996). ness. This study does not consider seedling establishment, Results of this study could explain why T. corym- which can affect the shape of the curves depicted in bosus does not ¯ower during the main course of the Fig. 9 (see also PicoÂ and Retana 2000). Seeds of T. summer, despite relatively high abundance of hum- corymbosus lack dormancy and are very prone to des- mingbirds in the forest (Amico 2000; Appendix E). iccation (personal observation; see also Calder [1983], Extrapolating from Fig. 5, I estimated that fruits de- Kuijt [1988]). Germinating immediately after dispersal, veloping from summer ¯owers would ripen during win- they infect the host within 1±2 wk. While a mean of ter when dispersal would be limited because the mar- 26% of all effectively dispersed seeds germinate, infect supial is inactive. However, peak ¯owering during fall the host, and survive the ®rst year until producing the as recorded in both study populations in 1997 and in ®rst true leaves (Amico 2000), the rate of seedling PenõÂnsula in 1998 may be somewhat maladaptive since establishment is not seasonally uniform. The propor- resulting fruit maturation peaks during December, 2 mo tion of dispersed seeds reaching the seedling stage in- before maximum fruit removal (Fig. 7). On the other creases by ϳ20% for fruits ripening during late Feb- hand, a higher proportion of ¯owers opened during ruary and March, at the end of the dry summer season August±September in Llao-Llao in 1998, resulting in (G. Amico, unpublished data). Because these mid- to a summer peak in fruit maturation that overlapped the October 2003 POLLINATION AND SEED DISPERSAL 2625

the peak in ¯ower opening occurs more commonly in the fall than in the winter. Lastly, fall-opening ¯owers could boost the pollination of winter ¯owers by en- couraging ongoing presence of this plant's hummingbird pollinator. However, this should lead to a highly asynchronous (and selectively unstable) onset of ¯owering because individuals starting ¯owering much later in the winter could increase their ®tness at the expense of early-¯owering individuals. One key question concerns the extent to which ¯owering time, duration, or intensity can change over evolutionary time (Rathcke and Lacey 1985, Fox and Kelly 1993, Wright and Calderon 1995). Responsiveness to selection on ¯owering phenology seems likely in the Loranthaceae, since southern South American species show a wide range of ¯owering phenologies (Hoffman 1982, 1995, Smith-RamõÂrez 1993, Rivera et al. 1996), suggesting that the family lacks intrinsic limitation on FIG. 9. Variation in per-¯ower contribution to ®tness over response to selection. The phenological behavior of T. the ¯owering period of Tristerix corymbosus (Eq. 2). Seasonal variation in mean per-¯ower contribution to ®tness was sta- corymbosus even varies over its geographical range. tistically signi®cant for the two study populations over the Although south-Andean populations do not ¯ower dur- two ¯owering phenologies: F9, 122 ϭ 5.16, P Ͻ 0.0001 for ing summer (see Riveros and Smith-RamõÂrez [1996]), Llao-Llao 1997; F9, 113 ϭ 5.61, P Ͻ 0.0001 for PenõÂnsula 1997; Chilean populations at a similar latitude near the Paci®c F9, 149 ϭ 19.07, P Ͻ 0.0001 for Llao-Llao 1998; and F8, 120 ϭ 4.41, P ϭ 0.0001 for PenõÂnsula 1998. Standard errors are not coast ¯ower year-around (Smith-RamõÂrez 1993). This shown for the sake of clarity. difference in phenological behavior can be consistent with the disperser-availability hypothesis if marsupial dispersers remain active during the milder coastal win- period of maximum disperser activity. Therefore, I pro- ter. Comparative studies on the relationship between pose the existence of strong selection for late-winter ¯owering phenology and seed dispersal in T. corym- ¯owering in T. corymbosus. bosus among these contrasting geographical localities Why does T. corymbosus not start ¯owering later or should be enlightening (see also Aizen et al. [2002]). concentrate ¯owering during the last months of winter? This study supports the hypothesis that ef®cient an- At this time, I can only suggest some possible answers. imal seed dispersers can become important agents of For example, use of temperature as an environmental natural selection on ¯owering phenologies of temperate cue triggering ¯owering may constrain evolutionary plant species. Flowers may open during suboptimal pe- possibilities. If ``temperature sum'' is the proximate riods for pollination if this ¯owering behavior increases trigger for ¯owering in this mistletoe (e.g., Diekmann seed dispersal. However, the proximate effect of tem- 1996), low winter temperatures would not be as effec- perature and other environmental factors on ¯ower and tive in triggering ¯owering. In addition, the strong, fruit development might curtail ®ne adjustment be- direct in¯uence of temperature on ¯ower longevity and tween ¯owering phenology and the seasonal activity rate of ¯ower opening may produce a ¯owering peak of animal mutualists. during fall. ACKNOWLEDGMENTS Another possibility is that I underestimated the long- I thank Peter Feinsinger, David Inouye, and Karen Searcy term ®tness contribution of fall ¯owers because the for their critical reading and useful comments on a previous 1997 and 1998 winters were unusually warm. In years version of the manuscript and Lyn Branch, Lawrence Harder, with cooler winters, slower fruit development and de- Bette Loiselle, Kerry Woods, and two anonymous reviewers layed ripening may increase the removal success for for their useful suggestions and careful editing. I am also grateful to the SecretarõÂa de Medio Ambiente of San Carlos fruits derived from fall ¯owers. However, the relative de Bariloche and the Pereda family for allowing me to conduct invariance observed in the ¯ower-level ®tness curves research in the Llao-Llao reserve and in the PenõÂnsula de San between 1997 and 1998 (Fig. 9), years showing an Pedro, respectively. The park ranger Alfredo ChaÂvez provided ϳ2ЊC difference in mean July temperature, suggests key logistical support, and Guillermo Amico provided skillful assistance during many freezing days of ®eldwork. This re- that the overall phenology of ®tness is basically un- search was funded by the Consejo Nacional de Investiga- altered over a wide range of winter temperature. ciones CientõÂ®ca y TecnoloÂgicas (PEI No. 0018/97) and the A third explanation is that fall ¯owers produced out- National Geographic Society (Grant 6192-98). side the optimum time are of little cost and represent LITERATURE CITED a bet-hedging strategy against complete reproductive Aizen, M. A. 2001. Flower sex ratio, pollinator abundance, failure in very severe winters (see also PicoÂ et al. and the seasonal pollination dynamics of a protandrous [2002]). However, this hypothesis cannot explain why plant. Ecology 82:127±144. 2626 MARCELO A. AIZEN Ecology, Vol. 84, No. 10

Aizen, M. A., and C. Ezcurra. 1998. High incidence of plant± linator service in Helleborus foetidus. American Journal of animal mutualisms in the woody ¯ora of the temperate Botany 88:1025±1032. forest of southern South America: biogeographical origin Hoffman, A. J. 1982. Flora Silvestre de Chile: zona austral. and ecological signi®cance. EcologõÂa Austral 8:217±236. FundacioÂn Claudio Gay, Santiago, Chile. Aizen, M. A., D. P. VaÂzquez, and C. Smith-RamõÂrez. 2002. Hoffman, A. J. 1995. Flora Silvestre de Chile: zona central. Historia natural de los mutualismos planta-animal del Third edition. FundacioÂn Claudio Gay, Santiago, Chile. Bosque Templado de SudameÂrica Austral. Revista Chilena Hopper, S. D. 1981. Honeyeaters and their winter food plants de Historia Natural 75:79±97. on granite rocks in the central wheatbelt of western Aus- Amico, G. 2000. DispersioÂn del quintral (Tristerix corym- tralia. Australian Wildlife Research 8:187±197. bosus) por el monito del monte (Dromiciops australis). Li- Horvitz, C. C., and D. W. Schemske. 1990. Spatiotemporal cenciatura thesis. Universidad Nacional del Comahue, Bar- variation in insect mutualists of a neotropical herb. Ecology iloche, Argentina. 71:1085±1097. Amico, G., and M. A. Aizen. 2000. Mistletoe seed dispersal Jaksic, F. 1997. EcologõÂa de los vertebrados de Chile. Edi- by a marsupial. Nature 408:929±930. ciones Universidad CatoÂlica de Chile, Santiago, Chile. Armesto, J. J., C. Smith-Ramirez, and C. Sabag. 1996. The Johnson, S. D. 1992. Climatic and phylogenetic determinants importance of plant±bird mutualisms in the temperate rain- of ¯owering seasonality in the Cape ¯ora. Journal of Ecol- forest of southern South America. Pages 248±265 in R. G. ogy 81:567±572. Lawford, P. B. Alaback, and E. Fuentes, editors. High lat- Kelt, D. A. 1994. The natural history of small mammals from itude rainforests and associated ecosystems of the west Aisen Province, southern Chile. Revista Chilena de His- coast of the Americas. Springer, New York, New York, toria Natural 67:183±207. USA. Kuijt, J. 1988. Revision of Tristerix (Loranthaceae). System- Barros, V., V. Cordon, C. Moyano, R. Mendez, L. Forquera, atic Botany Monographs 19:1±61. and O. Pizzio. 1983. Cartas de precipitacioÂn de la zona Lechowicz, M. J. 1995. Seasonality of ¯owering and fruiting oeste de las provincias de RõÂo Negro y NeuqueÂn. Facultad in temperate forest trees. Canadian Journal of Botany 73: de Ciencias Agrarias, Universidad Nacional del Comahue, 175±182. Cinco Saltos, Argentina. Martin, F. M. 1959. Staining and observing pollen tubes by Brody, A. K. 1997. Effects of pollinators, herbivores, and means of ¯uorescence. Stain Technology 34:436±437. seed predators on ¯owering phenology. Ecology 78:1624± Mazer, S. J., R. R. Nakamura, and M. L. Stanton. 1989. Sea- 1631. sonal changes in components of male and female repro- Bronstein, J. L. 1995. The plant±pollinator landscape. Pages ductive success in Raphanus sativus L. (Brassicaceae). 256±288 in L. Hansson, L. Fahrig, and G. Merriam, editors. Oecologia 81:345±353. Mosaic landscapes and ecological processes. Chapman and Motten, A. F. 1982. Autogamy and competition for polli- Hall, London, UK. nators in Hepatica americana (Ranunculaceae). American Calder, D. M. 1983. Mistletoes in focus: an introduction. Journal of Botany 69:1296±1305. Pages 1±17 in M. Calder and P. Bernhardt, editors. The O'Neil, P. 1999. Selection on ¯owering time: an adaptive biology of mistletoes. Academic Press, New York, New ®tness surface for nonexistent character combinations. York, USA. Ecology 80:806±820. Cocucci, A. A. 1991. KolibribluÈten der Subantarktis. Der PicoÂ, F. X., H. D. Kroon, and J. Retana. 2002. An extended Palmengarten 1/91:1±8. ¯owering and fruiting season has few demographic effects Diekmann, M. 1996. Relationship between ¯owering phe- in a Mediterranean perennial herb. Ecology 83:1991±2004. nology of perennial herbs and meteorological data in de- PicoÂ, F. X., and J. Retana. 2000. Temporal variation in the ciduous forests of Sweden. Canadian Journal of Botany 74: female components of reproductive success over the ex- 528±537. tended ¯owering season of a Mediterranean perennial herb. Endler, J. A. 1986. Natural selection in the wild. Princeton Oikos 89:485±492. University Press, Princeton, New Jersey, USA. Primack, R. B. 1985. Longevity of individual ¯owers. Annual English-Loeb, G. M., and R. Karban. 1992. Consequences Review of Ecology and Systematics 16:15±37. of variation in ¯owering phenology for seed head herbivory Primack, R. B. 1987. Relationships among ¯owers, fruits and and reproductive success in Erigeron glaucus (Composi- seeds. Annual Review of Ecology and Systematics 18:409± tae). Oecologia 89:588±595. 430. Feinsinger, P., H. M. Tiebout, and B. E. Young. 1991. Do Ramsey, M. 1995. Cause and consequences of seasonal var- tropical bird-pollinated plants exhibit density-dependent iation in pollen limitation of seed production in Blandfordia interactions? Field experiments. Ecology 72:1953±1963. grandi¯ora (Liliaceae). Oikos 73:49±58. Fox, G. A., and C. K. Kelly. 1993. Plant phenology: selection Rathcke, B. 1988. Interactions for pollination among co¯ow- and neutrality. Trends in Ecology and Evolution 8:34±35. ering shrubs. Ecology 69:446±457. Fox, J. F. 1992. Pollen limitation of reproductive effort in Rathcke, B. J. 1992. Nectar distributions, pollinator behavior, willows. Oecologia 90:283±287. and plant reproductive success. Pages 113±138 in M. D. GoÂmez, J. M. 1993. Phenotypic selection on ¯owering syn- Hunter, T. Ohgushi, and P. W. Price, editors. Effects of chrony in a high mountain plant, Hormathophylla spinosa resource distribution on animal±plant interactions. Aca- (Cruciferae). Journal of Ecology 81:605±613. demic Press, New York, New York, USA. Harder, L. D., and W. G. Wilson. 1994. Floral evolution and Rathcke, B., and E. P. Lacey. 1985. Phenological patterns of male reproductive success: optimal dispensing schedules terrestrial plants. Annual Review of Ecology and System- for pollen dispersal by animal-pollinated plants. Evolu- atics 16:179±214. tionary Ecology 8:542±559. Rau, J. R., D. R. Martinez, J. R. Low, and M. S. Tilleria. Herrera, C. M. 1988. Variation in mutualisms: the spatiotem- 1995. DepredacioÂn de zorros chillas (Pseudalopex griseus) poral mosaic of a pollinator assemblage. Biological Journal sobre micromamõÂferos cursoriales, escansoriales, y arbor- of the Linnean Society 35:95±125. Âcolasõ en un aÂrea silvestre protegida del sur de Chile. Re- Herrera, C. M., A. M. SaÂnchez-Lafuente, M. Medrano, J. vista Chilena de Historia Natural 68:333±340. GuitiaÂn, X. CerdaÂ, and P. Rey. 2001. Geographical varia- Reid, N. 1991. Coevolution of mistletoes and frugivorous tion in autonomous self-pollination levels unrelated to pol- birds? Australian Journal of Ecology 16:457±469. October 2003 POLLINATION AND SEED DISPERSAL 2627

Rice, W. R. 1989. Analyzing tables of statistical tests. Evo- Stone, G. N., P. Willmer, and J. A. Rowe. 1998. Partitioning lution 43:223±225. of pollinators during ¯owering in an African Acacia com- Rivera, L. G., L. Galetto, and L. Bernardello. 1996. Nectar munity. Ecology 79:2808±2827. secretion pattern, removal effects, and breeding system of Tadey, M., and M. A. Aizen. 2001. Why do ¯owers of a Ligaria cuneifolia. Canadian Journal of Botany 74:1996± hummingbird-pollinated mistletoe face down? Functional 2001. Ecology 15:782±790. Riveros, M., and C. Smith-RamõÂrez. 1996. Patrones de ¯or- Totland, é. 1994. Intraseasonal variation in pollination intensity and seed set in an alpine population of Ranunculus acioÂn y fructi®cacioÂn en bosques del sur de Chile. Pages acris in southwestern Norway. Ecography 17:159±165. 235±250 in J. J. Arroyo, C. Villagran, and M. Kalin Arroyo, Waser, N. M., L. Chittka, M. V. Price, N. M. Williams, and editors. EcologõÂa de los bosques nativos de Chile. Editorial J. Ollerton. 1996. Generalization in pollination systems, Universitaria, Santiago, Chile. and why it matters. Ecology 77:1043±1060. Schemske, D. W., M. F. Willson, M. N. Melampy, L. J. Miller, Wheelwright, N. T. 1985. Competition for dispersers, and the L. Verner, K. M. Schemske, and L. B. Best. 1978. Flow- timing of ¯owering and fruiting in a guild of tropical trees. ering ecology of some spring woodland herbs. Ecology 59: Oikos 44:465±477. 351±366. WideÂn, B. 1991. Phenotypic selection on ¯owering phenol- Smith-Ramirez, C. 1993. Los pica¯ores y su recurso ¯oral ogy in Senecio integrifolius, a perennial herb. Oikos 61: en el bosque templado de la isla de ChiloeÂ, Chile. Revista 205±215. Chilena de Historia Natural 66:65±73. Willson, M. F., and C. J. Whelan. 1993. Variation of dispersal Stephenson, A. G., T. C. Lau, M. Quesada, and J. A. Winsor. phenology in a bird-dispersed shrub, Cornus drumondii. 1992. Factors that affect pollen performance. Pages 119± Ecological Monographs 63:151±172. Wolfe, L. M., and J. L. Burns. 2001. A rare continual ¯ow- 136 in R. Wyatt, editor. Ecology and evolution of plant ering strategy and its in¯uence on offspring quality in a reproduction. Chapman and Hall, New York, New York, gynodioecious plant. American Journal of Botany 88: USA. 1419±1423. Stiles, F. G. 1977. Coadapted pollinators: the ¯owering sea- Wright, S. J., and O. Calderon. 1995. Phylogenetic patterns sons of hummingbird-pollinated plants in a tropical forest. among tropical ¯owering phenologies. Journal of Ecology Science 198:1177±1178. 83:937±948.

APPENDIX A A table presenting results of ANOVA testing of the effects of supplemental hand- vs. open pollination (treatment) and month of ¯ower opening on the mean number of pollen tubes per style and fruit set is available in ESA's Electronic Data Archive: Ecological Archives E084-068-A1.

APPENDIX B A table presenting the number (means Ϯ 1 SE) of pollen tubes and fruit set of open- and hand-pollinated ¯owers and fruit retention time of exposed vs. bagged fruits is available in ESA's Electronic Data Archive: Ecological Archives E084-068-A2.

APPENDIX C A table presenting results of ANOVA testing of the effects of month of fruit maturation on the proportion of fruits removed and bagging treatment (i.e., exposed vs. bagged fruits) and month of ¯ower opening on fruit retention is available in ESA's Electronic Data Archive: Ecological Archives E084-068-A3.

APPENDIX D A table presenting results of ANOVA testing of the effects of month of ¯ower opening on the proportion of fruits that were removed by animals and bagging treatment (i.e., exposed vs. bagged fruits) and month of ¯ower opening on fruit retention time is available in ESA's Electronic Data Archive: Ecological Archives E084-068-A4.

APPENDIX E A ®gure depicting seasonal changes in hummingbird visition to an arti®cial feeder and weather over a 4-yr period is available in ESA's Electronic Data Archive: Ecological Archives E084-068-A5.